Replace all occurrences and regular expression

Question

I'm trying replace two consecutive line breaks with the HTML tag <p/>. So in a string such as:

\r\n\r\n\r\n

There are two consecutive occurrences of \r\n\r\n,

The result should be:

<p/><p/>

but with C# String.Replace, it only detects the first occurrence and I just get back:

<p/>\r\n

So I'm wondering if any regular expression gurus know how to detect that using regular expression?

Edit:

I figured the question is a bit confusing. Let me rephrase it. The requirement should be to replace any "\r\n" with a tag <p/> only if there is another "\r\n" immediately before it.

Such with the string:

\r\n\r\n\r\n

• The first \r\n, does not have another \r\n before it, nothing should be done,
• The second \r\n, it does have another \r\n before it, qualifies for replacement,
• The third \r\n, it does have another \r\n before it, also qualifies for replace.

So the result should be:

<p/><p/>

Answer 1

Yes, you can do this with a regular expression:

string tidyString = Regex.Replace(originalString, @"(?<=\r\n)\r\n", "<p/>");

If performance is an issue then you might find that rebuilding the string manually is quicker than a regex, but the drawback will be the more complicated code. (I'd probably go for the regex in 99% of situations.)

var sb = new StringBuilder(originalString.Length);
int startIndex = 0, nextIndex;
while ((nextIndex = originalString.IndexOf("\r\n", startIndex)) >= 0)
{
    if ((nextIndex == startIndex) && (startIndex > 0))
        sb.Append("<p/>");
    else
        sb.Append(originalString, startIndex, nextIndex - startIndex + 2);

    startIndex = nextIndex + 2;
}
sb.Append(originalString, startIndex, originalString.Length - startIndex);

string tidyString = sb.ToString();

Answer 2

You have said this in a comment:

What I want to do is, if I detect a "\r\n" right before a "\r\n", it should be replace with the tag "p".

This can be rephrased:

What I want to do is, if I detect N consecutive occurrences of "\r\n" where N > 1, I want to replace these with N - 1 occurrences of the tag "p".

Don't believe me?

Let's say we have some fence posts:

|  |  |  |  |

If I say, "I want to detect wherever the is a fence post right before another fence post," what am I really saying? I'm saying I want to make N - 1 detections wherever there are N consecutive fence posts (again: when N > 1).

Look above; there are five fence posts, but four occurrences such as what you've described. The same holds true for detecting consecutive occurrences of "\r\n" after the first occurrence.

---

It sounds like rather than replace the string "\r\n\r\n" with "<p/>", you want to replace all occurrences of "\r\n" after the first occurrence with "<p/>", stripping the first occurrence out. (Your example does not contain two consecutive occurrences of "\r\n\r\n"; it contains three consecutive occurrences of only "\r\n".)

If this is exactly what you want, I'd recommend simply doing what I just described: detecting a "\r\n" and replacing subsequent "\r\n" occurrences in the string with "<p/>" (until reaching a block of text that is different).

Let me know if you need any help doing that. It's also possible I've misunderstood you, of course; but this is what it looks like to me you are trying to accomplish.

Hint: A very easy but inefficient way of doing this would involve calling string.Split with the string "\r\n" as the delimiter and using the resulting array to reconstruct the string.

Actually, I don't think the above hint will quite get you there, even as an inefficient solution. There are obviously other ways, though.